The development and the reproductive output of 26 Bombus terrestris colonies were investigated. Four important points in colony development are distinguished. These are: a) the start of egg laying by the queen, leading to the beginning of the eusocial phase, the emergence of the first workers; b) the moment which we term the switch point, at which the queen switches from laying diploid eggs (producing workers or queens) to the laying of haploid eggs (producing males); c) the onset of queen production reared from diploid eggs; d) the loss of dominance by the queen, expressed by the beginning of aggression on the part of queen and workers, worker oviposition, oophagy and the functional elimination of the queen. This we call the competition point. The onset of queen production is highly correlated with the competition point but not correlated with the switch point. The pattern in time of the egg laying by a queen differed from the rate of increase of worker numbers published by previous authors. After an initial and slow phase, in which two broods of egg cells are produced, the rate of egg laying of a queen becomes constant, whether or not the eggs are fertilized. The transformation from this rate of egg laying to the earlier descriptions of increase in number is only possible if one neglects differences between colonies in time spent in the production of the two broods as well as the highly variable time needed for development into adults. The time at which the competition point occurs is much more predictable (at day 30.8 ± 4.9 after the emergence of the first worker) than the switch point. The latter ranges from day 6 to day 32. In our colonies two groups can be discerned, one of early switching colonies (at day 9.8 ± 2.4), the other of late switching colonies (at day 23.4 ± 4.6). The occurrence of the switch cannot be predicted from preceding behavioural or demographic data. As a consequence of the early switch such colonies produce mainly males. Partly due to the unfavourable larva/worker ratio only a few queens were reared from the last laid diploid eggs. In contradiction the late switching colonies produce on average sexuals at a 1.3:1 sex ratio (1:1.7 investment ratio). Remarkable however, biomass of sexuals is equal in both types of colonies. Males are about half the weight of queens. Certain colonies invest in males only ("3 males : 0 queens"). Since the two types of colonies occurred in about equal numbers, our local population is characterized by a 4:1 1(♂:♀) sex ratio. This male biased sex ratio, also observed for other bumblebees species (B. terricola and B. melanopygus, OWEN et al., 1980; OWEN & PLOWRIGHT, 1982), contrasts with existing theoretical models based on kin selection (TRIVERS & HARE, 1976) or local mate competition (ALEXANDER & SHERMAN, 1977) arguments. A model is proposed in which the males of the early switching colonies monopolize the matings due to their early appearance in the field. In fact, the early switching colonies, by investing mainly in males, parasitize on the reproductive strategy of the late switching colonies, which are characterized by a 1:1 sex ratio. At population level the two reproductive strategies are in equilibrium at a 1:1 frequency.